The present invention relates to an image processing apparatus that removes periodicity, caused by an influence of a halftone dot pattern, from an image formed through a halftone dot processing.
When a size enlargement/reduction processing is applied to an image formed through a halftone dot processing, deterioration of the image quality is caused by the fact that its halftone dot pattern has been also enlarged or reduced. Accordingly, it has been performed to remove the periodicity, residing in the halftone dot pattern, from the image concerned, before applying an image processing, such as the size enlargement/reduction processing, etc.
However, when a smoothing processing employing a spatial filter, such as a Gaussian filter, etc., is applied to the image so as to remove the periodicity abovementioned, unless the size of halftone dot pattern (a number of halftone dot lines) applied to an image to be processed (hereinafter, also referred to as a processing objective image) is fitted to that of the Gaussian filter to be employed, the effect of removing the periodicity would become insufficient, or otherwise, edge portions of partial images, such as characters, etc., would be excessively smoothed.
To overcome the abovementioned shortcoming, for instance, Tokkai 2004-295318 (Japanese Non-Examined Patent Publication) sets forth the image processing apparatus that generates plural kinds of image data, each of which is processed by each of Gaussian filters having different sizes of halftone dot pattern, so as to calculate each of dispersion values for each of the processed image data, and then, specifies a specific Gaussian filter having an optimum size of halftone dot pattern, based on the dispersion values calculated in the above.
In this connection, one periodical area of a halftone dot pattern is liable to form a rhombus or a parallelogram. Accordingly, when a normal Gaussian filter, such as a 3×3 pixel filter, a 5×5 pixel filter, etc., is employed for this purpose, it has been quite difficult to achieve such a skilled filtering operation that makes it possible to accurately remove the periodicity of the halftone dot pattern while preventing the excessive smoothing.
On the other hand, in order to apply a spatial filter, shaped in a parallelogram being same as the shape of one periodical area of the halftone dot pattern, to the image data concerned, the calculation processing, required for convolution arithmetic calculations, etc., has become complicated and cumbersome, resulting in the deterioration of the processing velocity and the complication of the hardware configuration.